ARM had a slow start on its way to move beyond microcontrollers and enter the high performance market. ARM Ltd made the Cortex A9, their first out-of-order core, in 2007. Throughout the 2010s, they gradually made bigger, higher power, and higher performance cores. Pushing performance boundaries isn’t easy, but today, ARM’s cores can be a viable alternative to Intel and AMD’s offerings in the server market.
RISC-V started much later, but has seen faster growth. Berkerly’s BOOM core had grown into a sizeable out-of-order design by 2016. Now, SiFive’s P870 looks a lot like ARM’s Cortex X series in terms of reordering capacity, core width, and execution units. It might not be a match for ARM’s best, since the load/store queues look a bit small and vector execution throughput is a bit weak. But from looking at P870, SiFive’s ambitions are clear. They want a chunk of ARM’s pie.
RISC-V is getting better and better at a rapid pace. The software side of the story still has a long way to go, but that, too, is getting better. Exciting.
The software side of RISC-V is quite healthy given that the hardware side is just starting to get there now.
These RISC-V SBCs are coming with full distributions of Debian. So, Linux is in decent shape.
Haiku had a working RISC-V port before they had ARM ( pretty amazing ).
Even SerenityOS is getting a RISC-V port right now.
Given the recent Qualcomm Alliance, I would expect a phone SoC with Android in the not too distant future.
Windows may be a bit longer. Does that really matter though?
RISC-V will be strong in the SBC space first ( already starting to happen ). If RISC-V is able to achieve Raspberry Pi like price / performance with mainline Linux kernel support ( I think we will get that ), it will probably take the SBC market from ARM. Unlike ARM, we may see laptops come next with RISC-V ( alongside SBC ). We may even see servers before phones and tablets happen. Since ARM has not really taken hold in servers yet, I could the market skipping straight to RISC-V ( alongside x64 ). ARM will likely stay strong in mobile devices for a while. Other than those making their own silicon though ( like Apple ), what will be the advantage of ARM over RISC-V longer term?
I’d be ready to ditch ARM in a heartbeat if RISC-V were able to avoid all the bullshit development barriers and portability hell plaguing ARM hardware today!!! Seriously if they could just make RISC-V as easy to boot into as x86 is without 1) having to hack into it, 2) being stuck on a proprietary kernel that can’t be upgraded, then RISC-V will get my business!
I am sick of the status quo on ARM. I’m actually very concerned that after qualcom and others gets their hands on RISC-V, they’ll turn it into another proprietary hellscape with massive effort to support individual devices. Only having to implement one standard boot process for every device would be a huge advantage for RISC-V in the eyes of FOSS enthusiasts over ARM.
I completely agree with you and hope RISC-V does it better than ARM has.
So far, I have been encouraged to see chips from StarFive and Sipheed getting mainline kernel support as well as work on things like SBI and UEFI for RISC-V.